Natural hydrogen occurrences are thought to be sourced either by radiolysis, where hydrogen is formed by water breakdown induced by natural radioactivity, or by water-rock reactions, where water is reduced by the oxidation of Fe2+-bearing minerals.
Serpentinisation is a ubiquitous water-rock reaction wherever olivine occurs on Earth, and is a known source of hydrogen, for example associated with the Oman Ophiolite. Geochemical modelling shows it be a highly effective process at generating low oxygen activities and high hydrogen activities (potentially high enough to create a hydrogen gas phase), with conditions sufficiently reducing to generate metallic iron alloys in some environments. Experiments have demonstrated hydrogen generation. Other sources of hydrogen from water-rock reaction that have been suggested include siderite, biotite granite, peralkaline granite and iron formations. Siderite has the disadvantage of generating more CO2 than hydrogen. Biotite and peralkaline granites have been shown experimentally to generate hydrogen but will create less reducing conditions than serpentinisation and hence lead to lower hydrogen activities. Iron formations have also been shown experimentally to react with water creating hydrogen, but at significantly lower hydrogen activities (and higher oxygen activities) than any of the other sources. Their mineral assemblages buffer significantly higher oxygen activities (and hence lower hydrogen activities) than the other potential sources.
The rates of hydrogen generation by radiolysis and water-rock reaction differ by many orders of magnitude with radiolysis likely to require of the order of a billion years of more to generate a usable resource. In contrast, experimentally determined rates of serpentinisation are rapid (even compared to hydrocarbon formation) and while quantification of the rates of hydrogen formation from other sources have not been quantified, they too are likely to be fast. However, kinetic effects and disequilibrium mineral assemblages occur in experiments so caution should be used in interpreting results.
The unfocussed nature of radiolysis as well as its extremely slow rate means that it is not readily considered in any hydrogen play. Similarly, the rarity of peralkaline granites makes them a low priority in exploration, and siderite’s co-production of CO2 renders it a less attractive exploration target. Serpentinisation of ultrabasic rocks and magnetite-bearing iron-formations remain as potential sources, with the former’s chemical ‘effectiveness’ (potentially forming a gas phase at depth) balancing perhaps the wider known occurrences of iron formations.